# Reception voids with two omni antennas at 90 degrees

We are using two omni ferrite rod antennas to transmit a signal at 125 KHz. The two antennas are positioned at 90 degrees from each other on the PCB plain, and the device installation is with the PCB parallel to the ground. The device will typically be installed on the ceiling, so we'd like to have good reception in a half-sphere around the device, to the sides and towards the floor.

The mobile receiver also uses two ferrite antennas that are at 90 degrees from each other. Naturally the orientation of the receiver with respect to the transmitter can change.

Overall reception of the 125 KHz signal is good, at about 5-6 meters, but we have some voids. In these voids reception drops to around 0.5-1.5 meters.

These voids are at roughly 45 degrees to the transmitter antennas, but we haven't mapped them all yet:

Can anyone give some clues as to why we would have these voids? In theory, having two transmitter omni antennas should cover all voids, no? Perhaps the angle between the antennas should be changed?

Edit: the transmission signal is wired in series to the two antennas. On the receiver side the receiver is wired separately to both antennas (i.e. not in parallel).

Edit 2: two more observations:

1. Disabling one of the receiver's antenna seem to improve reception.
2. On the other side of the 45° void (i.e. 180° from it) reception is better, almost as good as in the rest of the field.

It depends on how the two antennas are wired. In general, if you're feeding two antennas with the same signal, you're not going to get "omni" coverage, you're just going to shift the directionality to a new angle. For example, a single loopstick has a figure-8 pattern, with nulls in two directions. Two loopsticks wired in parallel will still have a figure-8 pattern, but the nulls will point in different directions.

The same thing applies on the receive side. If you wire two antennas to a single receiver, you're just going to shift the directionality. To get true omni coverage, you need two separate receiver front-ends.

If you can phase-shift the signal to/from one of the antennas by 90° relative to the other, you can create a "circular" polarization pattern that will help mitigate this problem. Now the nulls will be in the plane of the PCB.

• On the receiver we have separate inputs from each antenna. The transmitter's antennas are connected in series. Does this improve signal coverage over in parallel? Apr 10, 2014 at 13:26
• No. Either series or parallel, both antennas are getting the same in-phase signal. Apr 10, 2014 at 13:34
• Two more observations (updated the question as well): disabling one of the receiver's antennas seem to improve reception, and on the other side of the 45° void (i.e. 180° to it) the reception is almost as good as the rest of the field. Can this shed some more light? Apr 10, 2014 at 14:03
• This is a well known characteristic in wireless communication. When signals of equal amplitude and opposite phase, a null exists called Rician Fading corrected by choosing the 1 of 2 antenna with the stronger signal in mobile receivers. When used to advantage to block certain directions, it is called Spacial Diversity.
– user38637
Apr 10, 2014 at 14:31
• @Joe - indeed, our receiver does just that, choosing the stronger signal from its two antennas. Apr 13, 2014 at 8:31

You can regard your two transmit coils as just one coil with maximum flux density at both +45 degrees and +225 degrees (when wired in parallel or series aiding). If wired in opposition max flux density is at -45 degrees and +135 degrees. It's just one coil you have effectively so why not make a loop coil in the plane of the ceiling and get full coverage?

EDIT